What you’re describing here is _good customer service_, which basically doesn’t scale at all ever (at least in any automated way). You provide more tools to expand the capability of CS, but at some point, you simply just have to hire more and more CS people.

This isn’t a knock by the way. Good customer service makes a customer feel like there’s someone just waiting at the company to help them. Kudos to you for providing that experience.

So while these aren’t great examples of “doing something that doesn’t scale (with the eventual goal of scaling)”, it’s a great example of “doing something that doesn’t scale (to provide a better product)”.

You can also give the tools directly to the customers, and you can bundle/package the tools into future versions of the product. A common pattern is to have a public git repo and the customers can git pull to get the latest versions of the support tools.

Shipping tools as a git repository sounds delightful, at least to my engineering-minded heart. So painless! Free versioning! Rollbacks!

If we were in the devtools business I would certainly consider that. Sadly, though, owners of small retail businesses tend not to be as comfortable with version control as one might hope…

I disagree: if you're a good product owner, it will aid with the goal of scaling.

For example, when you're sitting there walking them through something they could do alone, are you just playing back some script, or are you getting an understanding of where they went "aha!" and making note of how you can embed that aha moment into the product itself?

You’ve found me out! Not only is it excellent customer service, it’s also priceless market research that I would be lost without. Side conversations during calls like that have led to more improvements and features (and in one case an entirely new side product) than I can possibly count.

Based on every mega-scale consumer tech company, the best way to scale customer service is to make it as useless and frustrating as possible.

Apple would be a notable exception. As would Amazon.

I think you’re referring specifically to Google and Facebook, and you’re right. I think the big difference is the amount of profit per customer.

Google and Facebook make a small amount from a large number of people so it makes sense that they could not profitably scale customer support. In the instances where they make a large amount of money off of a small number of people, mainly ad sales, their support is much better.

I at least don't expect anything from FB.

I must also say Uber's CS is absolute garbage. I just cannot believe how bad they are. I have been unable to use uber for weeks and have been reaching out to them non-stop every few days. It is like speaking to a brick wall.

Have you tried social media? Companies can be strangely responsive on those channels.

Try reaching out to their Facebook or Twitter/X profiles.

I don't use X but might try facebook.

I've had consistently bad experiences with Apple support.

Even amazon seems to do a lot to err on the side of "automatically make the customer happy" rather than "actually have someone look into the problem" (I did a recent round of "order a three-pack, get a singleton labelled as a three pack, return for replacement, the replacement was also a singleton, return for refund" - they were prompt about the replacement, and prompt about the refund, very straightforward - but I actually wanted 6 of the items (and got someone else to not even order them to avoid dealing with this.) This is a pretty straightforward inventory problem, and I'm not convinced amazon even got any "signal" about it from the interaction. (And after all, the alternate items I ordered... were still from amazon, so they don't have a problem either, really...)

I had this same issue a few time too. Ordered five 3-packs. Receive one three pack, and 4 singles.

I am convinced a robot just snags "one of each" from a bin, and if some packs are broken apart and separate... then they meet that logic.

> Even amazon seems to do a lot to err on the side of "automatically make the customer happy"

I can't find the article now, but I read a piece about the volume of returns and what happens to them ("reverse logistics"). Basically, companies find it's more profitable to always allow returns and keep the customers happy -- and often, they'll let the customer keep the "returned" product because even if the product works 100%, the cost to ship and verify is below what they'll make reselling it.

I thought this was an interesting contrast to LL Bean, who a few years back decided to end their lifetime guarantee, and naturally upset some lifetime customers.

Apple and Amazon aren't event remotely in the same league. Amazon chat agents use canned responses (probably due to being required to handle multiple customers at one time.) I've had one tech support/customer service interaction by phone with Apple and it was great - I needed my out of warranty MBP battery replaced. The agent was kind, knowledgeable, focused and took her time working with me to diagnose/confirm the issue and setup the return (they sent me a box & label to send the MBP in for service.) I had my MBP back within 5 business days from the day they picked it up. Apple is expensive, but they take phone support seriously.

And then there are companies like Dell and Lenovo which dispatch a serviceman to your address on next day and fix or replace the device on the spot.

Isn’t that normally an add on service that you purchase in addition to the thing you’re buying?

In my experience, they do not even ask for warranty or support service or what not. A Lenovo person will come and help. Period.

Dell ProSupport for IT is the absolute benchmark

My ipod 3rd gen broke in 2005. Mailed support, I had a new one mailed few days later. They didn't even bother checking whether it was broke, or even retired it, they just sent me a new one.

I don't like Apple too much and try to avoid their products, but can't lie never had such a good customer experience.

Probably depends what you need. I’ve used Amazon quite a lot for quite a long time and I’ve never had an issue go unresolved. I guess I’ve also almost never had to talk to a human though, they have pretty well automated just doing whatever the customer wants them to in most cases.

The last time I had to talk to a human (and it was two years ago) was because I accidentally left a fire stick behind in Puerto Rico two years prior and someone bought a few seasons of True Blood with my Amazon credentials. They still refunded me, even though it took me two years to notice the charge.

If Amazon delivers a late package I ask for a $5 credit, and I get it every time

Apple support seems to come and go in waves depending on when the last time someone looked at the expenses were. In my time working there, we'd have months where everything was by the book and the book was the Apple Care service contract written up by the lawyers, or where every customer got the hard sell on all the metrics (they didn't have many but those they had we needed to push).

We also had months where one could get assigned to spending 3 whole days helping a customer port all of their emails and photos out of a corrupted windows outlook installation over to their new mac for no other reason than they'd asked for help getting it done.

Or where you would be celebrated for spending 5 hours with a 65 year old grandmother buying her first computer carefully going over all her needs and specs and down selling her from what the competitors had convinced her she needed (seriously, no 65 year old grandmother getting her first computer needs the hardware or associated software to cut together professional film in order digitize her collection of VHS home movies) and finish up with a sale of maybe $1300 and directions to competitor down the street because they have better photo printers than the few we carry in stock all on the (IMO correct) theory that a customer who can trust you to exactly what they need and nothing more and send them to a competitor for a better product is one who will come back again and again on the basis of that trust.

I don't know how expensive those two interactions were in the short term, but I do at least know that 65 year old grandmother became a regular customer and could not stop telling all the other customers how much she loved us. For as many problems as I did have with how Apple hamstrung their people, I do wish more companies were even half as good as that.

Amazon's support is absolutely amazing compared to a company like Uber.

Apple is only good compared to the total garbage fire of MS/Google, and Amazon's support has been horrible for years now. If they don't have an automated answer to your problem you're SOL.I had them ship me the wrong item with the correct item's SKU taped on it and they argued with me for quite awhile before just refunding me.

MSFT is pretty good with their hardware support.

I believe that for the surface line, but I mainly deal with the O365 folks.

> If they don't have an automated answer to your problem you're SOL

No? I've had issues with an external monitor and had an applecare specialist debug a bunch of settings and then schedule a call with me for the next day after they heard back from the escalation team. Thankfully I bought a different cable which fixed my issue, but they are very nice for support from my experience.

I can't think of any instances where I've needed Google or Facebook support. I pay Google money for premium services, so I'd hope they'd be better, but I've read plenty of evidence to contrary.

I was thinking more specifically of the Ubers, Lyfts, DoorDashes, Instacarts of the world. They are providing a consumer service, don't really control the end-to-end experience that well, and when things go wrong they are very difficult to work with on a reasonable resolution.

I'm pretty sure that's literally what they do. The metric for customer service at scale is "engagement" where bigger numbers are worse, so they hellban you to a network of automated answering services, online documentation and chatbots which all redirect you to each other while avoiding any question that doesn't have a quick, easy, do-it-yourself solution.

You're not their customer. You're their product.

If I'm ordering groceries on Instacart, how am I not their customer? Am I only the customer of the grocery store and delivery driver?

To be frank, I’ve never used Instacart, so anything would be (a hopefully educated) guess.

But it might be some version of Michael Seibels's classic "Who is your customer?" (https://youtu.be/C27RVio2rOs?t=10m11s)

A more subtle aspect of customer service here is that, as the dev or PM responding to the customer, you have a lot more power to give the customer what they want.

A BigCo can hire a lot of CS people but the best they can do sometimes is "we hear you and we'll pass along your feedback".

Alas, you’ve found the reason this won’t scale. Doing customer support as the CTO is a superpower (up to a point!) for both user growth and product design. But there’s going be to come a point where we have to hand some portion of support over to a dedicated support team, and no matter how well we train those folks they’re just not going to be quite as empowered and effective. The longer I can kick that can down the road, though, the better!

(At least for the business. My sleep schedule would improve amazingly!)

You want Support Engineers who are up to speed on your code review processes and standards and given access to commit "straightforward" bug fixes. They exist :)

https://gitlab.com/gitlab-org/gitlab/-/merge_requests/125249 https://gitlab.com/gitlab-org/gitlab/-/merge_requests/131316 https://gitlab.com/gitlab-org/gitlab/-/merge_requests/131469 https://gitlab.com/gitlab-org/gitlab/-/merge_requests/130988 https://gitlab.com/gitlab-org/gitlab/-/merge_requests/132806...

From there, you need to build a culture that is welcoming of "strangers" contributing code. You get these two things, you get nits and gotchas fixed directly from pain points customers are having, while product engineering is (mostly) focusing on feature dev.

This is simple support tiers.

Tier 0 are effectively secretaries who can file structured info around issue.

Tier 1 are dedicated support people who can follow scenarios and guide customers through the product.

Tier 2 are what you call "support engineers". They know the product, features, code, upcoming features and so on. For an in house product they are capable of making straightforward bugfixes.

Tier 3 is sometimes called "vendor support". For an in-house product this is effectively product development team.

As you can see, good supports bleeds into or blends with product development. This is how you get support answers like "this feature is planned to go live Y24Q1, but you can sign up to beta in exchange for feedback" or at least "This is not supported, but you can use features x and y to achieve similar result", instead of "Sorry, such workflow is not supported"

Doing things manually for a while might be natural to some builders/hackers. Customer service is one such example where many builders are willing to go above and beyond. So maybe the OP does not have to read too much into the PG quote.

I think where it helps with scaling is that they’re very involved in all the pain points. None are hidden by client silence. Then they can build this learning into the product. Simplify, fix UI, add, remove. That digs you out of the daily effort trap.

> which basically doesn’t scale at all ever

It basically scales fine, the problem for companies trying to bump margin is it's roughly linear with customers...

> Investigate problems and repair live data, usually just by logging in as a user and using the tools in the product. Can they do this themselves? Sure. Do they appreciate our doing it instead? Oh, hell yes.

This is definitely one that doesn’t scale as eventually they don’t allow devs to see prod data, so using user accounts is a no go at my org for that reason.

Makes things infuriating for both clients and customer support.

Not being allowed to see production data (or even in some cases see production!) was my greatest frustration in my last corporate job. But to be fair, keeping engineers off the live data was…pretty non-negotiable at most customers.

Which leads to some pretty boring days onsite, when “onsite” means “in the SCIF”

> SCIF

Though there's a big sliding scale there, between "not allowed to look at prod user data for privacy/etc reasons" and "the giant black hole of nothing-comes-out" in a SCIF.

If “devs” can’t have access to production data then they can’t be responsible for production issues, simple as.

I would say 95% of my intellectual horsepower at work goes into queries and analysis of production data tables, event streams, and logs to investigate issues or confirm key invariants are holding. Production is infinitely more creative than anyone sitting down to write unit tests. Organizations refusing to learn from it, as a matter of policy, either have incredible confidence in their testing and formal verification regimes… or much more likely don’t give a shit about correctness.